Data Visualization
What to Know
Data visualization questions test whether you can choose an appropriate chart, encode data honestly, and design dashboards that answer a question. Core concepts: chart selection by data type and intent (bar for categorical comparison, line for trends over time, scatter for correlation between two measures, histogram for distribution, box plot for spread and outliers, heatmap for density across two dimensions). Understand perceptual accuracy: position and length are read more precisely than angle, area, or color saturation. Know the parts of a chart (axes, scales, legends, annotations) and the difference between a static report and an interactive dashboard with filters and drill-downs.
Traps
Watch for truncated y-axes that exaggerate differences, dual axes that imply correlation, pie charts with many slices or slices that do not sum to a meaningful whole, and 3D effects that distort area. Another trap is confusing correlation with causation when a scatter shows a pattern. Questions may ask which chart best shows part-to-whole (stacked bar or pie with few categories), distribution (histogram), or ranking (sorted bar). Color traps include using red/green alone without accessibility alternatives and rainbow scales for ordered data. Dashboard traps: too many metrics on one screen, inconsistent time filters across tiles, and mixing aggregated and row-level values.
Worked Example
You must show monthly revenue for five regions over two years. A line chart with one line per region and a shared zero-based y-axis is appropriate; sorting or stacking would obscure trends. Rule of thumb: if the x-axis is time, default to a line; if comparing categories, default to a sorted bar; if showing two measures' relationship, use a scatter. Always start quantitative axes at zero for bars, and label units explicitly.